This K07 proposal in cancer prevention, control, and population sciences will enable the applicant to build on our existing knowledge of individual-level pathways and trajectories of tobacco use by examining contextual factors and group level processes to yield more comprehensive understandings of the etiology of tobacco use. Peer groups play a critical role in the initiation and progression of adolescent smoking. Most studies of youth smoking have used the individual as the unit of analysis and peer smoking has been measured by adolescent report of their friends' smoking. The conceptual framework and methodology of social network analysis utilizes peer-nomination data to construct friendship links that allow patterns of relationships among individuals to be determined and interpreted within the context of the larger social environment. This methodology has been underutilized as a mechanism for understanding the development of health behaviors in youth and few studies exist that have used the peer group as the unit of analysis to understand social influences and youth tobacco use. The research design comprises two studies, each aimed at understanding the social context in which peer group membership is related to youth smoking. Study 1 will use existing data from the National Longitudinal Study of Adolescent Health to examine whether influence on smoking occurs within the context of adolescent peer groups across two time points. Potential moderator variables such as gender, ethnicity, and parental variables will be examined. Study 2 will involve two years of novel data collection that will focus on vulnerabilities to prosmoking socialization influences in a sample of pre-adolescents in one social network system of an ethnically diverse urban middle school. The goals of the career development plan are to broaden conceptual and analytic knowledge of social network analysis as it can be applied to peer group formation and stages of cigarette smoking in youth and to develop innovative methodologies to bridge the analysis of individual and group level data. This training plan is intended to provide mastery of skills in group level analyses and promising new methodology that will enable the applicant to develop an independent program of cancer prevention research. The long-term objectives are to translate knowledge derived from studies of contextual influences on the etiology of tobacco use into culturally sensitive, developmentally appropriate, and novel prevention interventions that reinforce protective influences in youth's social environments.